You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The following code will compile, but cause the runtime error because inferred type is Number which is unsupported. val b = mk.ndarray(mk[1.0, 2]). Considering current library design, is it possible to move this check to compile time? Maybe numbers should be treated as Double?
The text was updated successfully, but these errors were encountered:
The mk[...] construct is simply a list. While it's possible to convert the types to a common type, it's currently a substantial effort for a minor benefit. This is because we'd need to first determine the larger type and then convert the remaining types to it, all while optimizing the process. I already have a task on my plate to create a parser for external data reading. Once I add that, it should address and close this issue as well.
It's worth considering whether this approach is the right choice at all. Since it exhibits non-intuitive behavior, users may be surprised by the high memory consumption.
And for now, I agree that it might be a good idea to add validation and throw a runtime exception
The following code will compile, but cause the runtime error because inferred type is Number which is unsupported.
val b = mk.ndarray(mk[1.0, 2])
. Considering current library design, is it possible to move this check to compile time? Maybe numbers should be treated as Double?The text was updated successfully, but these errors were encountered: